One missing thing: a dual mult. You need to send the same signals (input and modulation) to the VCAs to get a smooth L-R panning action. But there's an issue here in how the panning itself would sound...
Stereo panners are set up (usually) with what's called an "Equal Power" crossfade. What this means is that as each audio signal approaches the center of the stereo platform, it starts to drop so that when the signal is at dead center, each VCA would be outputting a 3 dB lower signal than the actual peak...that, in truth, would be from the extreme hard-panned direction to about 35-40% of the way to center. But you need to do this because if both VCAs were outputting an unattenuated signal, the center 10-20% of the stereo field would be abnormally loud. By lowering the VCA outputs at each VCA approaches the center, the summed level comes out to be pretty much the same as a single unattenuated VCA's. The result is a smooth motion across the stereo field without a big over-level "lump" in the middle. This is also how you'd want to do this if the VCAs were panning a control signal; again, you don't want this "zone" in the middle of the pan where the CV levels go waaaaay up, as that same, smooth equal power crossfade is also useful in keeping control signals from driving other modules into difficult to control areas.
Now, as to how to do that...well, that depends on the VCA in question, plus your own patching methods.